Volume 10 -
Evidence Based Health Policy, Management & Economics 2026, 10 - : 42-68 |
Back to browse issues page
Download citation:
BibTeX | RIS | EndNote | Medlars | ProCite | Reference Manager | RefWorks
Send citation to:
BibTeX | RIS | EndNote | Medlars | ProCite | Reference Manager | RefWorks
Send citation to:
Abbasian-Naghneh S. Cost-Effectiveness of AI-Assisted Diagnostic Decision Support: Systematic Review. Evidence Based Health Policy, Management & Economics 2026; 10 :42-68
URL: http://jebhpme.ssu.ac.ir/article-1-569-en.html
URL: http://jebhpme.ssu.ac.ir/article-1-569-en.html
Department of Finance, Na.C., Islamic Azad University, Najafabad, Iran , salman_abasian@iau.ir
Abstract: (23 Views)
Introduction: Artificial Intelligence (AI)-assisted Clinical Decision Support Systems (CDSS) are increasingly integrated into diagnostic care, but their economic value remains uncertain. This systematic review aimed to evaluate the cost-effectiveness of AI-assisted CDSS in diagnostic settings and identify factors influencing their economic outcomes.
Materials and Methods: Primary economic evaluations of AI-assisted CDSS used for diagnosis, screening, or clinical triage were identified through searches of PubMed/MEDLINE, Embase, Scopus, Web of Science, the Cochrane Library, the NHS Economic Evaluation Database archive, and Google Scholar up to 15 June 2026. Reporting was guided by PRISMA 2020. Methodological appraisal used the Philips checklist for model-based evaluations and the Consensus Health Economic Criteria for empirical evaluations, while reporting quality was assessed using CHEERS 2022. A separate formal risk-of-bias assessment was not conducted. Study selection, data extraction, and appraisal were performed by one reviewer, with subsequent verification. Because of substantial clinical and methodological heterogeneity, findings were synthesized narratively. The review protocol was not prospectively registered.
Results: Nine primary economic evaluations published between 2021 and 2026 were included, covering cancer screening, ophthalmological screening, multispecialty diagnostic support, and emergency triage. Five studies identified at least one economically dominant AI-assisted strategy. Favorable economic outcomes were most frequently associated with reduced diagnostic workload, replacement of existing workflow components, and prevention of downstream healthcare costs. Other interventions increased expenditure without sufficient clinical benefit. Economic outcomes were sensitive to AI acquisition and maintenance costs, diagnostic performance, disease prevalence, implementation strategy, and healthcare perspective.
Conclusion: AI-assisted CDSS may be cost-effective in selected diagnostic settings, but economic value depends on clinical context and implementation conditions. The evidence remains limited by heterogeneous study designs, predominantly model-based evaluations, restricted real-world validation, and single-reviewer assessment. The absence of prospective protocol registration and complete original screening records further limits methodological transparency.
Materials and Methods: Primary economic evaluations of AI-assisted CDSS used for diagnosis, screening, or clinical triage were identified through searches of PubMed/MEDLINE, Embase, Scopus, Web of Science, the Cochrane Library, the NHS Economic Evaluation Database archive, and Google Scholar up to 15 June 2026. Reporting was guided by PRISMA 2020. Methodological appraisal used the Philips checklist for model-based evaluations and the Consensus Health Economic Criteria for empirical evaluations, while reporting quality was assessed using CHEERS 2022. A separate formal risk-of-bias assessment was not conducted. Study selection, data extraction, and appraisal were performed by one reviewer, with subsequent verification. Because of substantial clinical and methodological heterogeneity, findings were synthesized narratively. The review protocol was not prospectively registered.
Results: Nine primary economic evaluations published between 2021 and 2026 were included, covering cancer screening, ophthalmological screening, multispecialty diagnostic support, and emergency triage. Five studies identified at least one economically dominant AI-assisted strategy. Favorable economic outcomes were most frequently associated with reduced diagnostic workload, replacement of existing workflow components, and prevention of downstream healthcare costs. Other interventions increased expenditure without sufficient clinical benefit. Economic outcomes were sensitive to AI acquisition and maintenance costs, diagnostic performance, disease prevalence, implementation strategy, and healthcare perspective.
Conclusion: AI-assisted CDSS may be cost-effective in selected diagnostic settings, but economic value depends on clinical context and implementation conditions. The evidence remains limited by heterogeneous study designs, predominantly model-based evaluations, restricted real-world validation, and single-reviewer assessment. The absence of prospective protocol registration and complete original screening records further limits methodological transparency.
Keywords: Artificial Intelligence, Biomedical, Clinical, Cost-Benefit Analysis, Decision Support Systems, Technology Assessment
Type of Study: Review Article |
Subject:
Health Economics
Received: 2026/06/25 | Revised: 2026/09/2 | Accepted: 2026/09/18
Received: 2026/06/25 | Revised: 2026/09/2 | Accepted: 2026/09/18
Send email to the article author
| Rights and permissions | |
 |
This work is licensed under a Creative Commons Attribution 4.0 International License (CC BY 4.0). |
